Lab

XPool AI - Supercluster Memory Optimizing Solution

Sesterce’s Xpool AI project is an initiative that utilizes the team’s expertise in microchip technology to enhance AI training on GPUs. This project uses a sophisticated base command manager tool to optimize GPU memory through advanced algorithms like page attention and offers a unique feature allowing users to rent GPUs at no cost. This approach increases the efficiency of AI training processes. With Xpool AI, Sesterce works towards the democratization of AI technology, making it accessible for a wide range of users and, in turn, transforms the landscape of AI technology.

XWATT - Containerized AI Cluster With Integrated Heat Reuse solution

Xwatt is an innovative containerized solution for AI training clusters and emphasizes environmental responsibility in the tech industry. This 100% green solution uses Direct Liquid Cooling (DLC) systems, which increases hardware performance and reuses the heat generated. The container model provides flexibility, meaning it can be easily transported and adapted to various locations—making it the ideal choice for organizations seeking mobile and adaptable infrastructures. Xwatt provides companies with a sustainable yet powerful solution for their AI projects.

ArX Usecase - Advanced AI computing with responsible energy practices

The ArX initiative centers around developing a European-made asynchronous inference chip for AI. The ArX chip is designed to efficiently handle diverse workloads, enhance infrastructure performance, and reduce energy consumption.

Sesterce Vision 2030

Sesterce plans to construct a specialized datacenter, tailored for Generative AI companies' model training requirements. This facility will use 100% DLC systems therefore eliminating traditional air cooling methods. This new approach will repurpose the excess energy generated to heat the surrounding buildings and infrastructures.

Sesterce Research Lab